Feeding the Machine

[Music]

Hello, I'm Alan Kohler, Editor-at-Large of Intelligent Investor and a Finance Presenter and Columnist and Podcaster for the ABC.

And I'm James Thomson, Senior Chanticleer Columnist at The Australian Financial Review.

And we are The Money Café. G'day, James.

How are you, Alan?

I'm very well, had a long weekend... It's a good thing the Australian market was closed yesterday because it clucked down on Friday in Wall Street, didn't it?

It did, yeah, a sudden reversal - well, two things happened, the US jobs data came in very hot, surprisingly hot, which had everybody rushing to redo their calculations for US interest rates...

It was the old 'good news is bad news' gag.

Exactly. Now, would you believe, we've gone from expecting a couple of cuts at the start of the year, we're now pricing in a hike for US interest rates before the end of the year, 0.25 per cent, because the economy's just running so hot.

Which is interesting because they've got a new Chairman of the Fed now and he's about to have his first meeting, Kevin Warsh, and the assumption has always been that he's Donald Trump's man, he'll do whatever Donald Trump says and Donald Trump wants rate cuts but the market is saying he's going to hike. That's interesting.

Yeah. What's really interesting, is that he might not want to hike himself, but he might be outvoted. He might really have no choice because there's no doubt that Donald Trump's running the economy hot, you've got the combination of tax cuts, deregulation and still pretty steady fiscal spending and then this is being juiced by big massive household equity wealth, which according to Bank of America, I think has gone up about $16 trillion dollars in the last 18 months. No wonder people in America are feeling rich and continuing to spend... Then, of course, you've got the energy shock which gasoline prices over there up around $4.50 US a gallon, that hasn't slowed the economy yet but it's going to at some stage, or it's certainly going to put upwards pressure on inflation.

The other thing is, all this borrowing, all this spending on AI, how does that go in a high interest rate environment? We did see on Friday night, as you say, tech stocks and AI chip stocks specifically sold off and what's interesting, Alan, is the biggest falls haven't been on Wall Street, but in Korea where a couple of chip makers have done very well, Samsung and a company called SK Hynix. The Korean market's up 90 per cent year to date, which is an incredible run, but it's had a couple of very bad days on Monday, while we were all off for the King's Birthday, the Korean market opened more than 8 per cent down and had to be shut for - it's called hitting its limits and had to be shut temporarily while they restored order, still finished 8.3 per cent lower. Wall Street hung on okay on Monday night and Korea will probably bounce on Tuesday. There's a lot of AI jitters, I think, around the world at the minute.

You had a piece yesterday saying, "We just saw how this bubble could end, blame Trump, Musk and Korea..." I guess what you were talking about was the impact of interest rates and how that could bring the bubble to an end, is that what you were talking about?

Well, I think that's one of the things. I think the other thing though, Alan, obviously we've been hearing a lot about the SpaceX float and IPO. I think my other thing is that AI could sort of eat itself to a certain extent. These companies are out there raising so much money, $75 billion dollars is being raised in the SpaceX float, we saw Google last week raising $80 billion dollars US. Anthropic's going to come to the market. Does it raise somewhere between $100 billion and US $80 billion? How much money in the world is around for these raisings, particularly when a lot of them - SpaceX is a perfect example - is mainly hot air? This is a deeply loss making company with prospects that might be great over 5, 10, 15, 20, 25 years... This is a great test of how willing we are to bet on this AI story and I think we might see some doubts.

Yeah, totally. I think all of those three big floats this year, SpaceX, Anthropic and OpenAI, they're going to be dragging a ton of money out of the market and they're going to go into the indexes, I guess, because the rules have been bent to make sure they do and that's going to mean that all these ETFs and fund managers have to buy them, which means they have to sell something else.

The good thing about that, Alan, is S&P which has the largest index, they've held the line and they won't let SpaceX in for the first 12 months while it's not profitable, so I think that's really important, that's a good sign. But yeah, I think you're right, there's something to AI and we've got a few questions on this later, everybody's spending money on it, nobody really knows what they're going to get out of it though and everyone's still very much still experimenting. These experiments aren't free, there's trillions of dollars being thrown at this in the hope that it works. We're not certain yet and so we sort of wait to see what it does and how it changes the economy at scale. This is not a sure thing.

I suppose a lot of this depends on what happens with inflation because if inflation continues to worsen, partly as a result of what's going on in the Persian Gulf, then the interest rates will have to be increased and will be increased by more than once or twice. So how do you see the inflation situation?

To me, there's no question that inflation is going to be higher for longer. AI is inflationary in this early stage because everyone around the world is trying to build data centres using the same equipment, everyone's rushing for the same equipment, the same people. There's these wonderful stories of electricians being flown around the world in private jets to finish off data centres and having spoken to some of the big data centre operators here, they're true and they say quite simply, there's no doubt that sticking an electrician on a private jet to get them to the other side of the world to beat a data centre delivery deadline, maybe that costs you $50,000 or $100,000. It's worth it because the penalties and obligations that you've got to the data centre users are so huge that you're happy to eat that cost. This is a fully fledged bubble that is inflationary to start with.

The world's gone mad, hasn't it?

Yeah and I think we have this combination of the re-tooling of the economy for AI and reshoring and friendshoring and all that stuff, defence spending is going up obviously. That's all inflationary, throw in the energy crisis, that is inflationary, whether it's temporary or not is the question. So I think we're in a world of higher inflation, higher rates and that makes returns tougher. None of that is reflected in, particularly US share markets. It's reflected in the Australian share market to a certain extent now, but the Australian share market is sitting on 17.5 times earnings, well above the long-term average of 14 times. What do you look at in the Australian economy and think, geez, that's fantastic? Commodities, I guess, because they're going into the AI build out. But there's not too much about the domestic economy that screams, this is a moment where we should be trading above our long-term averages.

Fair enough. The other thing that's kind of supporting the market in a way, is the fact that the oil price hasn't gone up as much as anyone thought it would. The international energy agency says it's the greatest disruption to oil and energy supplies in history, yet the oil price is under $100 US dollars a barrel, didn't go to $150 and looks okay. You, in a piece yesterday, called it a 'creepy calm'. Why creepy?

I think that's right - that was a piece by my colleague, Anthony, but I think he's got it exactly right. This feels weird. Not only is Iran not sorted out, it doesn't look like it's even close to being sorted out and even if we get the sort of white flags and the all clear, it will take months to rebuild the infrastructure and get oil flowing again, but yet we've muddled through really well. What appears, is that that muddling has been a result of two things, America increasing its oil exports and China basically stopping oil imports and using up its strategic reserves, at much lower prices which is good for China, but it's also been very good for the rest of the world. How long are those two buffers going to hold? Nobody really knows.

As Jeff Currie, he's a former Goldmans Commodity Strategist, keeps saying, "You can print money, but you can't print molecules of oil and gas." So is there like this moment coming where the numbers just don't work anymore, the physical supplies are sort of exhausted? I don't know... Again, we've heard lots of doom and gloom predictions about that, we were supposed to reach that moment in March and then April and then May and now I see some more predictions this morning that we're going to hit this sort of moment of reality in June, well who knows? Everyone's still flying to Europe, from what I can see. It's got to get tight at some time, doesn't it?

Yeah and in the meantime, it's terribly complicated in Iran or the Middle East. Trump clearly has to accept some humiliation at some point because Iran simply isn't giving in, isn't agreeing to - I mean, it's continuing to demand cash, it wants its frozen assets back, it wants to keep control of the Strait of Hormuz, all these things that basically Trump doesn't want to concede, but he might have to. The other complication that's current at the moment is what Israel does, both in Lebanon and Iran. Trump says he controls what he calls the shots about Israel, but I'm not sure that's true.

The Iranians are no one's friends, a murderous regime, it's oppressed its people forever, but they've sort of outplayed Trump at his own game. 'Mr Art Of The Deal', he's failed to see the leverage that Iran's built up and used pretty cleverly. You just go back to the very start of this and you sort of think, what did Trump expect was going to happen? I know he's not a great one for listening to people, but surely there were some people inside the military White House establishment saying, "This isn't going to work in three days like in Venezuela." Anyway, here we are three months later.

Before we go to questions, can you give us a quick summary of the KPMG story, what's going on there?

KPMG is obviously one of the big four accounting consulting professional services firms in this country. They're in the middle of a scandal, it's already claimed their CEO after, inside their audit team they've fessed up to the fact that members of the audit team were using information from one audit of big listed companies to try and pitch for and try to win the audit roles at other big listed companies.

How were they using it?

That's a good question. There was confidential data. From what we can glean, it looks like they were sort of saying, "Well, this is how it works over here, so this is what we can do for you over there..." KPMG is very clear that one client didn't get to see the other client's data, but the use of the data even to try and pitch for clients is a big no-no, and obviously the audit is a statutory part of what public companies do and private companies and it's sort of a bit holier than thou and KPMG's really done it this time with this scandal.

And there was a whistleblower, wasn't there?

A whistleblower raised these problems, the whistleblower is still unidentified, he's been given an apology from KPMG and treated very poorly by the firm which has done several investigations...

What a surprise!

...Done several investigations into their claims and again, what a surprise, originally found there was nothing to see here. So now it looks like that investigation was a whitewash and there's more investigations going. It's one of these rolling messes that will just keep rolling and it will claim some other scalps, I think, over time.

Do you think it's going to change the way people do an audit? I suppose the ultimate would be if these firms, all of them, had to split off their audit function from their professional services. Because there's always been questions about the potential for conflict of interest.

There has been and there's been suggestions that that should happen and then those recommendations go to the Government and again, surprise, surprise, nothing happens because these big partnerships are hard to reorganise. Will this be the scandal that breaks the donkey's back and sparks some reform? I don't know, probably should be but it doesn't always work like that.

Well, the firms don't want to split off their audit function from the professional services because they do use audits as the foot in the door, right?

That's right.

Which is arguably a conflict to begin with.

Yes, that's true and that's why the split has been suggested for a long time. Alan, you sort of see this one, you see the PwC tax leak scandal a few years ago, there's all sorts of calls for reform and again, nothing happens. I'm not convinced that something's going to happen this time either.

I wonder why nothing happens, I suppose these outfits are fairly well plugged in, are they?

They're well plugged in, most of them work for the Government in some sort of consulting role. I guess they're big partnerships, they're unusual sort of organisations and they're sort of big enough to be companies but they're not companies and I think some of the partnership structures are quite arcane, you unpick one thing, you've got to fix another bunch of things. It seems to get chucked in the too-hard basket, unfortunately.

Okay, before we get to questions, here's a quick word from our sponsor.

[Recording]

Tax, housing and investing could be about to change in a big way.

Following the proposed changes to capital gains tax and negative gearing in the federal budget, Alan Kohler and Intelligent Investor founder John Addis are hosting a live webinar to unpack what it could mean for investors.

They'll cover housing affordability, property investing, shares, ETFs, retirement planning, and the risks and opportunities for younger Australians.

Join The Rules Are Changing: Tax, Housing & Your Wealth, live on Thursday 18 June at 1pm AEST.

Register now at https://www.intelligentinvestor.com.au. Places are limited.

[End Recording]

Just a general advice warning, this is general advice only and not personal advice. If you need personal advice, please go and see an adviser. Karen says, "I enjoy Alan's point that houses are meant to have just one job, they're for living in and the personal wealth vehicle solutions need to be elsewhere, so why in Australia is the primary residence still the only place for tax free capital growth? To that, I wonder about tools like the individual savings accounts, as are found in the UK. British ISAs are their tax-free way for people to hold and build medium-term wealth, they allow various kinds of responsible personal investments towards securing one's personal future and so on, so why don't we have them here?" Do you know that?

Well, I guess we have super, would be the answer, Alan, which is not tax-free, but fairly tax advantageous system of retirement savings and I think that's what the primary reason for the ISAs are, they are retirement savings...

Do you know what the rules are with ISAs?

No, I don't.

Because we have investment bonds here which are 30 per cent taxed if you hold them for 10 years, I think we've got a question or two on those things too.

Yeah.

So we've got super and investment bonds, so we have something equivalent. I don't think they have those things, they just have ISAs.

Yeah and look, it's an interesting idea, Karen. What I hear a lot of as you go around the world, we obviously get to talk to lots of visiting people, often countries have come up with ideas that try to replicate or mimic super and you speak to a lot of people from those countries, including Britain and the US, they would just much rather the holistic program of super that we've got here. But I take Karen's point, it's not a medium-term solution, is it?

Peter says, "Morning, just wanted to give kudos to Stephen for giving us some rational critique of the BS being pulled in the SpaceX IPO..." and I think we've talked a bit about this, Alan, but Peter makes the point that SpaceX is going to be, basically, a very big bet on AI more than space and there seems to be no effort to read how Grok is an AI basket-case - Grok is part of the SpaceX family and just to sum up, Peter's basically saying that the SpaceX float is going to be a disaster and well done to Stephen for pointing it out.

I think Grok's market share of the AI space is 1.4 per cent, versus OpenAI and Anthropic's shares of above 30 per cent each, so it's well and truly in the rear.

Yeah.

Okay, Ben says, "I have a question about cash deposits and the financial claims scheme which guarantees deposits of $250,000 per account holder, per institution. As I understand it, small institutions like credit funds offer higher interest rates on savings than the big banks because they don't have the same ability to obtain low cost finance themselves, probably also a higher risk to depositors. If the FCS guarantees the deposits, why just not go for the highest rate on offer and throw caution to the wind?" Well, it's your money, Ben. If you want to throw caution to the wind, go ahead! That's the good thing about this sort of thing, there's choice, isn't there?

Yeah, that's right, I'm not sure how high the deposit rates on offer at small institutions are? I haven't seen any stonkingly higher than larger institutions but maybe Ben's seen something that he thinks he can take advantage of, which is fair enough.

Paul says, "No one begrudges Australia's lowest paid workers receiving a 4.75 per cent pay increase, but the Australian Chamber of Commerce and Industry' (ACCI) estimates the economy-wide wage bill impact at more than $11.7 billion. Even if only half of that is passed onto consumers, more than 30 per cent of the cost flows straight back to Government through additional tax income, GST, payroll tax and workers compensation premium. So here's the question, if the aim is to lift real wages, why does Government get such a big cut? Business pays, consumers fund it, workers keep only part of it and Government quietly clips the ticket. Wouldn't a cleaner policy be to pair minimum wage increases with targeted tax relief for low paid workers so more of the rise actually lands in workers pockets?"

Well, the tax paid on the minimum wage, which is $52,254.80, is 14 per cent, if you take into account the tax-free threshold and the 16 per cent rate. The Government gets 14 per cent of that $11.7 billion dollars which is $1.6 billion. I think the targeted tax relief for low paid workers already exists, doesn't it, with the tax-free threshold?

Yeah and I guess the other thing is, what Paul is suggesting, is a fairly command and control type economy there, where you're controlling wages and prices in a way. The other thing is, not everybody necessarily spends that, do they, Alan? Although, I guess if it comes through in wages, the Government does take a chunk, so that's true.

Yeah, well that's right, that's the whole point of it. I mean, obviously it flows through to other people than those on minimum wages and they tend to get an average of 30 per cent, that's true, but that's the thing. I mean, we've said before that we think too much reliance in the tax system on income taxes and there should be more on assets and consumption, but that's a tax reform too far, apparently, for the Government.

Indeed.

Blake says, "No one asked for my tax advice..." Blake's got a tax system reform suggestion, "Increase the tax-free threshold to $75,000, then 45 per cent after that. Every worker is better off..." and so on. I get an email once a week from somebody with a whole new tax reform, tax package, tax system suggestion and they're usually kind of theoretically fine, but never going to happen, so we move on.

There's a few good ideas around that. Jeff says, "Long-time listener, first-time questioner..." Thanks, Jeff. "If two people earn the same economic income but pay very different tax because one can structure it through companies, trusts, capital gains or super, is that really fair? Should tax systems focus on eliminating arbitrage by taxing income consistently regardless of form and let progressivity come from explicit payments instead of complex tax rules?"

Well, I'm not quite sure what he's talking about, but if he's talking about income splitting through trusts, the Government's trying to deal with that by having a minimum tax rate on trust income and a minimum tax of 30 per cent on capital gains. Super obviously is taxed less, 15 per cent. It's interesting that they are putting 30 per cent on a whole bunch of things and they seem to be trying to establish a minimum tax on everything other than employment income of 30 per cent, which I guess has been kind of established, hasn't it?

Yeah, well, as long as it gets through the parliament.

Yeah, that's right.

Yeah, that's an interesting way to look at it - well, sorry, I think you're right in response to Jeff, that there is an element of equalising all that around it, although the person who tries to restructure their affairs and get paid through companies taxes, capital gains or super, there's all sorts of small but important differences in that. Super's a good example, like if you're 33, then structuring affairs to get more money through super doesn't benefit you for the best part of 40 years. There's all sorts of wrinkles to this when you start trying to structure your affairs for tax, isn't there?

That's right. Harry says, "I've been catching up on the podcast, keep up the good work. A couple of weeks ago, Alan and James touched on themes including the lack of AI contingencies in the budget and Australia's relatively low consumption tax. How might you suggest Australia implement a specific consumer tax on AI tokens and how much above the current 10 per cent GST?" That's quite a long question, so I won't read it all, but it's an interesting question about AI tax, because people have been suggesting that in order to deal with the expected possible decline in income taxes as jobs get replaced by AI, we need to tax AI and specifically people have been suggesting a tax on AI tokens. But it seems to me that - and Harry kind of alludes to this - GST should apply to AI, right? I mean, it's a service that's provided and it should get GST.

I subscribe to something called Perplexity, which is an AI service and the subscription is $200 bucks a year, which divides into $16.67 a month and they're billing me - this is $16.67 US, $200 US - and they're billing me for some reason, I don't know quite why, but they're billing me $16.50 US a month, which turns into $23 bucks Australia, but that $16.50 does not include GST, right? So, they are not paying GST on my AI subscription. I asked Perplexity, what's going on with GST on AI and AI tokens and it was a very convoluted answer and the answer is, basically, it's not happening. It should be happening, but kind of it's not.

The Tax Office rules clearly do apply to AI tokens, so I reckon a start would be, make it happen.

We obviously had at the Financial Review, an AI summit last week and this suggestion was, Daniel Petre, who's a veteran venture capitalist, philanthropist, he was on stage and he said, "I'm a believer in the Bill Gates view of the world, we're just going to have to tax the robots." But you get to the heart of it, I've been thinking about it since then and how do you tax the robots, is a really interesting question. Do you tax token usage through GST or other methods? Is there some sort of super profits tax on AI firms? But how would you define them? Is Microsoft an AI firm or is Anthropic an AI firm or are they both AI firms? How do you make sure that those companies, largely US multinationals, actually pay their fair share of tax? Which we've never really been able to do in a good way.

Well, that's it, a decent start would surely be to apply a 10 per cent GST to AI tokens, which is what should happen, because it's a foreign service and they're basically a product, it's being sold in Australia and GST should apply, but it is clearly not applying. Harry suggests, "What about having an AI specific usage tax of 30 per cent?" Yeah, okay, fine, let's start with 10 per cent GST.

Yeah and again, what are we taxing?

Well, tokens are clearly a thing, AI tokens are created and then sold to consumers.

Yes, but a usage tax would be different to that again, wouldn't it? I mean, the GST would sort of be the best way of doing it.

Absolutely.

It's a sales tax rather than a usage tax. I mean, this is the...

I see, well, the usage - I don't know...

This is the complication. AI sort of asks you to rebuild the entire economy in a way that we can't even change the capital gains tax rules without starting...

James, if the SpaceX prospectus is correct, then GST on what they're doing would be enough to fund the entire Government, you wouldn't need any other tax. Obviously, the SpaceX prospectus is a fantasy and it isn't going to be like that, but there's no doubt that there's a lot of money now going on AI and it's going to increase a lot.

I'm just thinking this through as we speak, Alan. Say you whack a GST on AI tokens, that means you're paying the GST, is that designed to discourage you from using AI tokens or is an AI tax more designed to redistribute the wealth created by AI?

I don't know, I'm paying GST on my hamburgers and it's not entirely discouraging from eating hamburgers.

Do you know what I mean? If what we're concerned about is taxing AI in some way to rebalance the economy, then taxing the users of AI, is that what we want to do, or do we want to tax the people making profits from AI, or both?

Well, profits are supposed to be taxed, they are taxed, unless they kind of base themselves in the Cook Islands or somewhere and that's another whole topic. But, yeah, profits are supposed to be taxed and then the consumption of the product that they make is also supposed to be taxed 10 per cent GST, so the question is not whether there should be a tax on this stuff, the question is whether you can collect it? Because there is a tax on it, they've just got to bloody well collect it.

Yeah, I think that's the point. Ray has a question on the SpaceX IPO which I think is worth indulging, Alan, because they make the point - one of the things they say, "I guess super funds in Australia will be all in." That is a really interesting question. This is a challenging one for super funds because there's a lot of risk and a lot of potential in a business like SpaceX. It's going to be a big part of the market. I think this is going to be really tricky for the Australian super funds. Their investment styles have become a lot more passive in that they don't take bets on individual companies as much as they take bets on indexes and broader portfolios. But I think Ray's hit on something here, this is a tough one of the super fund industries, as well Anthropic and OpenAI be.

There's an argument that all of these AI firms, eventually what Anthropic, OpenAI and SpaceX are doing or trying to do, will become commoditised. They're all training off the same data, there's only so much data in the world to train off. They'll essentially be commoditised services where one of them will emerge as the winner. That's one theory. That's not a great thing to bet on for long-term wealth, which is what super is supposed to do, so I think Ray's right, this is a really interesting thing for super funds.

If you were betting on AI, you probably wouldn't bet on Grok, would you? I mean, you'd probably bet on Claude, I would have thought. If anyone's going to be the Google of AI, it'll be Claude, surely.

Maybe. My hesitation, Alan, is that I think about six months ago we thought OpenAI was unstoppable.

That's true, that's a fair point.

Leadership can change pretty quickly.

I might skip to Mark, who says, "About three weeks ago in your research and diversions, you provided a link to this article about two court rulings in China..." I should add that research and diversions is a section in my Weekend Briefing for Intelligent Investor that basically is a bunch of links that I find to various things and there was a link to it, a story about China courts preventing a company from laying someone off because of AI. They tried to reduce his salary by 40 per cent to begin with because AI was doing his job and then when he refused to accept the 40 per cent pay cut, he got laid off. He went through the courts in China and all of the levels of court from the Magistrates up to the Federal Court supported him and told the company they couldn't lay him off.

Mark says, "Do you think Australia should legislate something like this? How practical do you think it is to think this kind of thing would be for us here in Australia?" And I think, Mark, you need to just bear in mind that that was the courts effectively using non-AI specific laws basically about unfair dismissal and it's an interesting question as to whether Australia's unfair dismissal laws will apply, can apply, to people getting laid off for AI. It's interesting because the Chinese courts basically saying that the jobs going to be done for AI is not a sufficient reason under their unfair dismissal laws to lay someone off. Obviously, this hasn't appeared in front of a judge in the Federal Court in Australia yet, but I wonder how that's going to work?

Yeah, it will be fascinating to see. What's interesting, Alan, is the Fair Work Commission is drowning in a different type of AI noise at the moment, it's been inundated by applications, a lot of which are around unfair dismissal, but they've been written with the assistance of AI, pages and pages and pages and it's basically clogging up the Fair Work Commission where they're struggling to deal with everything. AI gets you one way or the other, I guess.

Yeah, they need to use AI to work through them. Claude will fix it for them. [Laughs]

Yeah.

Your turn.

I'm going to jump to Phil, who says, "Alan, you've suggested quarantining the shelving of the 50 per cent discount to housing only and leaving it available for shares. If that were adopted, couldn't I just set up a small company to buy a house and then sell the shares to get the 50 per cent discount?" I'm not sure about that, Phil. That sounds pretty complicated to me, Alan.

I reckon the Tax Office might regard that as a tax avoidance scheme, possibly.

Maybe, yes. But Phil also has a specific question for me, "Off the back of the AFR AI Summit and dealing with executives and sit in leaders who are advanced in their thinking of AI, did I get the sense that any of them were suffering from AI psychosis, which is when you basically think that AI's going to be much better than it is? In the US, across tech firms, there are lots of anecdotes of leaders being convinced that AI could solve a problem because they don't fully understand it end to end. They mock something up in AI, think it automates a process without seeing all the work people have to put in to finish it. I've heard of senior leaders disparaging a team's effort, saying they could have vibe-coded a solution." Vibe-coding is where you just basically say to AI, "Can you come up with this?" "I wanted to know if James got that vibe from senior leaders at the forum, if they had a disconnect from the promise of AI to the actual benefits?"

Phil, I think there's a touch of AI psychosis about the entire world at the moment. What's really interesting to me, I think Australian executives are actually going fairly slowly on AI at the moment. Peter Tonagh, who's the Chairman of Nine Entertainment, which publishes The Fin, he's also the Executive Chairman of Woolworths owned business called Quantium, it's a big data and AI consultancy. He says, what he gets told from the big AI firms is Australia's second in the world for adoption and then tenth in the world for actual deployment.

I think we're going fairly slowly here, not too much AI psychosis, but what I would say is that my sense is that nobody thinks they can afford to dally on this because they are scared that one of their competitors will come out, whether that competitor's in Australia or around the world, will come out with a sort of holistic AI solution that allows them to rewire their business and allows them to compete and bring prices down in a way that the Australian business won't be able to compete with. I think that fear factor, that sort of natural law of competition is very much working in the favour of the AI companies because nobody is willing to take the risk that this isn't something real, if that makes sense.

Well, it's definitely something real. How can anyone conclude that it isn't? Did you find that people at the AI summit are saying, "Maybe this is just a fad..."?

No, not that, Alan, but what they are seeing is that in the last six months or so, we've moved to this agentic AI where AI is being asked to solve very complex problems. Guess what, that chews a lot of tokens up. I think Alan's explained on this podcast before that 75 tokens is 100 words, right? Agentic AI, the average is that as an AI agent, before it even answers a question, just starts to understand it, it'll eat up 96,000 tokens. It's the same as the amount of text in The Great Gatsby, that's just to start thinking about a query. That increased token usage means much higher costs, so I think there is a little bit at the moment in Australia and around the world of people saying, "Hang on, just run that by me again, it's costing this much and what am I getting out the other end?"

You're right, companies are getting very lumpy invoices and you're right, they're all going, "Crikey, what's going on here? This is much more expensive than we thought."

Yes and they're not seeing the sort of productivity boom at the moment. Again, this is the thing about AI, everybody says, "Oh, that'll come. Costs will go down and that will come." But you look at what's happening in the AI infrastructure build out, the costs of a data centre are soaring exponentially. We're up to US $100 billion dollars for the cost of a 1 gigawatt data centre. Costs aren't coming down, they're going up.

Yeah, I agree. I do think you're right, there is a crunch coming at some point where the willingness to pay is going to collide with the difficulty of AI companies meeting the prices that people are prepared to pay, because they can't, the cost of producing this AI is so great.

I think Australians are being fairly circumspect on all that, they're trying to weigh it all up. But again, Alan, this is the thing, you walk away from a summit like this and - agentic AI is six months old, six months. The speed at which this is moving and that we're going from, 'Does it work?', to, 'Okay, what's the cost?', is so quick. The next twist and turn is just around the corner. Finish us off, Alan.

Last question and I wanted to do the one about childcare, it's a bit of a long question but I think it's really worth having. "I watched your recent ABC segment on declining birth rates and felt like there's a missing piece, the sheer reality of under-five children childcare costs. My partner and I are in our late 30s with a two-year-old child and a baby in Sydney's East. On paper, our combined income puts us in the top 5 per cent of household earners, we pay heavy taxes, massive mortgage, attempting to keep our careers growing, stable, with very little family support, so it's a juggle. Yet, after paying top bracket income tax and a jaw-dropping $100,000 dollars in after-tax dollars for daycare, virtually no subsidy, the maths just doesn't math.

We live in a two-bedroom apartment and we have a very small disposable income after insurances, utilities, groceries, etcetera... Everyone I know in our scenario is living a very modest life despite apparent amazing salaries. To rub salt into the tinned tuna dinner, we have a very material sum locked away in super, great for our 60s, but frustratingly untouchable while we're stretching our every dollar for a young family in our 30s. Our generation was told that if we worked hard, kept both careers moving, we'd be set up. Instead, women are returning to high pressure roles purely to keep their career equity alive, essentially working for zero net return today until public primary school kicks in.

The resulting stress is taking a massive toll on the wellbeing of our young families. Is it any wonder our demographic is opting out because it looks like financial madness and a way to kill your wellbeing? My questions for the Café, what structural reforms could the childcare system introduce that incentivises people to pursue a successful stimulating career while choosing to have children? And two, what are the long-term social and economic ramifications for Australia if our biggest group of tax paying professionals stop reproducing, potentially creating a bifurcated population?" I think it's so interesting and really, the other thing that's going on is the rise of One Nation. A lot of people, particularly in regional areas are just saying, "We've had enough of this."

Yeah, you would have this too, Alan, you write about intergenerational inequality and you get a cohort of people, they're almost always male, saying, "What intergenerational inequality? This is a made up idea, people just need to put their heads down and work harder, that's what I did." I think this email just blows that to smithereens.

It's real, no doubt about it. Anonymous specifically asks, "What structural reforms could the childcare system introduce?" And he brackets, "I'd love to see childcare paid for pre-tax dollars." The number one thing would be to say that childcare is a cost of earning your income, so therefore it's tax deductible. I think that's the number one thing you could do. In my article recently on childcare, I said it should be made free like public education after five. This couple's hanging out until they can send their kids to primary school after five so they don't have to pay, apart from whatever it is, $2,000 a year for books and everything. But under five, it's expensive. Make it free! That would require nationalisation, because it's no good 100 per cent subsidising a private sector industry, that's pointless.

Yeah and as for the long-term social and economic ramifications for Australia, I think this is a real issue. Demographic changes, the sort of K-shaped economy where the older asset rich households have much greater spending power than younger less asset rich households is a real issue. I think the history and what we're seeing around the world says this doesn't end well. Polarised societies, tend to elect populist leaders, populist leaders tend to spend lots of money keeping inflation and interest rates higher and just pouring fuel on the same fire. This is why we do need to find ways to tackle this intergenerational equality, because otherwise you get a divided society that doesn't work for anyone. I think this is a great email that goes to the heart of that.

Okay, thanks everyone for listening to today's Money Café, I'll be back next week with Stephen Mayne, send in your questions to themoneycafe@intelligentinvestor.com.au. Until then, I'm Alan Kohler, Editor-at-Large of Intelligent Investor and the Finance Presenter and Columnist and Podcaster for the ABC.

And I'm James Thomson, Senior Chanticleer Columnist at The Australian Financial Review.

[Music]

Got a question for next week? Please send it to themoneycafe@intelligentinvestor.com.au.